Study characteristics
The initial literature search yielded 1397 articles from all the databases. A final total of 20 studies13–32 were included after screening based on the inclusion criteria (Fig. 1). All of the studies had a total of 2,990 participants which included 684 MIS-C patients. Except for five studies20–22, 29,31, rest of the studies had fewer than 100 enrolled participants. Eight studies13–20 compared MIS-C and COVID-19 along with subgroup analysis of MIS-C and severe/non-severe COVID-19; five studies13,14,21,22,24 compared MIS-C and KD, five studies21–23, 25,26 compared severe and non-severe MIS-C, three studies21,23,27 compared MIS-C with and without coronary artery abnormality, four studies22,30−32 compared KD during and before SARS-CoV-2 epidemic, while two studies28,29 compared MIS-C across different age periods (0–4/0–5 years representing the younger age of infants or preschoolers, 5–12/6–12 years representing the medium age of school-age and 13–20 years representing older age of puberty or post-puberty). All study features and characteristics are presented in Table 1. In the quality assessment of study design for the selected studies, three studies14,16,31 were deemed of moderate quality, with the scores of 6, and the remaining 17 studies were deemed of high quality, with scores above 7. (eTable in the Supplement)
Meta-analysis of inflammatory markers (Table 2, eFigure 2–7)
MIS-C vs. severe/non-severe COVID-19 13–20
We found no statistically significant difference in white blood cell count (WBC) or leukocytes (× 109/L) [weighted mean deviations (WMD) (95%CI): 0.14 (-0.91, 1.18), p = 0.799], procalcitonin (PCT) [standard mean differences (SMD) (95%CI): 0.05(-0.55, 0.66), p = 0.864] and ferritin [SMD (95%CI): 0.27(-0.10, 0.65), p = 0.157] levels. However, platelet count (PLT) (× 109/L) levels were overall lower in MIS-C patients than COVID-19 patients [WMD (95%CI): -102.48 (-122.74, -82.22), p < 0.001], as observed in the fixed-effects model.
For other inflammatory markers, due to the presence of significant heterogeneity of outcomes, subgroup analyses were performed for MIS-C vs. COVID-19, based on severity of COVID-19. Pediatric patients of COVID-19 not meeting the criteria for MIS-C were categorized into: severe COVID-19 (children with ARDS, or requiring invasive respiratory support, or an increase in positive pressure ventilation above the baseline) and non-severe COVID-19 (children with minimal symptoms or asymptomatic, without requiring invasive respiratory support). Compared to non-severe COVID-19 patients, MIS-C patients had lower absolute lymphocyte count (ALC) (× 106/L)levels [WMD (95%CI): -1481.71 (-1978.78, -984.63), p < 0.001], while MIS-C patients had higher levels for absolute neutrophil count (ANC) (× 106/L) [WMD (95%CI): 2678.74 (1673.14, 3684.35), p < 0.001], C-reactive protein (CRP) [SMD (95%CI): 1.38 (0.89, 1.86), p < 0.001] and D-dimer [SMD (95%CI): 1.56 (0.95, 2.18), p < 0.001]. Compared to severe COVID-19 patients, MIS-C patients had lower levels for lactate dehydrogenase (LDH) [SMD (95%CI): -0.91 (-1.60, -0.22), p < 0.05] and higher levels of erythrocyte sedimentation rate (ESR) (mm/hr) [WMD (95%CI): 37.87 (14.66, 61.09), p < 0.01], while same levels of ALC(× 106/L), ANC(× 106/L), CRP and D-dimer [WMD (95%CI): -39.29 (-655.52, 576.95), WMD (95%CI): -5708.41 (-17274.50, 5857.68), SMD (95%CI): 0.08 (-0.33, 0.50), SMD (95%CI): -0.25 (-0.82, 0.31), p > 0.05].
A moderate degree of heterogeneity was reported in four comparisons: ALC in MIS-C vs. COVID-19(p < 0.05, I2 = 71.4%), ALC in MIS-C vs. non-severe COVID-19(p = 0.043, I2 = 59.4%), ANC in MIS-C vs. severe COVID-19(p = 0.063, I2 = 71.0%) and CRP in MIS-C vs. COVID-19(p < 0.05, I2 = 69.5%). Rest of the comparisons showed no significant difference in statistical heterogeneity.
MIS-C vs. KD13,14,21,22,24
Compared to patients with KD, MIS-C patients had lower levels for ALC(× 109/L) [WMD (95%CI): -2.07 (-2.34, -1.79), p < 0.001] and PLT(× 109/L) [WMD (95%CI): -219.46 (-237.41, -201.51), p < 0.001], while higher levels were noted for CRP [SMD (95%CI): 0.57 (0.25, 0.90), p < 0.01] and ferritin(ng/mL) [WMD (95%CI): 647.00 (464.91, 829.09), p < 0.001] and the same level was noted for ESR(mm/hr) [WMD (95%CI): -21.33 (-61.91, 19.26), p = 0.303]. Only one comparison of CRP level showed a moderate degree of heterogeneity (p = 0.069, I2 = 51.2%). Rest of the comparisons showed no significant differences in statistical heterogeneity.
Severe MIS-C vs. non-severe MIS-C21–23, 25,26
Severe MIS-C patients had higher levels of WBC (× 109/L), CRP (mg/L), D-dimer (µg/mL) and ferritin (ng/mL) [WMD (95%CI): 5.41 (1.67, 9.14), 76.30 (49.29, 103.31), 2.46 (1.72, 3.20), 351.64 (118.62, 584.67), p < 0.01] compared to non-severe MIS-C patients. For levels of ALC (× 109/L), ANC (× 109/L), PLT (× 109/L), and fibrinogen (mg/dL) [WMD (95%CI): -0.32 (-0.65, 0.00), 4.21 (-1.05, 9.47), -17.38 (-38.76, 4.00), -0.01 (-1.59, 1.58), p > 0.05], there were no significant differences between both the groups. Two comparisons, ANC (p = 0.065, I2 = 70.7%) and fibrinogen (p = 0.074, I2 = 68.8%) levels showed moderate degree of heterogeneity. Rest of the comparisons showed no significant differences in statistical heterogeneity.
MIS-C with coronary artery abnormality vs. MIS-C without coronary artery abnormality21,23,27
There were no statistically significant differences in ALC (× 109/L) [WMD (95%CI): 0.15 (-0.19, 0.49), p = 0.400], PLT(× 109/L)[WMD (95%CI): -12.45 (-39.70, 14.81), p = 0.371], CRP(mg/L)[WMD (95%CI): -14.98 (-89.28, 59.31), p = 0.145], D-dimer (µg/mL) [WMD (95%CI): -0.99 (-2.75, 0.76), p = 0.268] and ferritin (ng/mL) [WMD (95%CI): 121.11 (-991.64, 1233.87), p = 0.956] levels among MIS-C patients with and without coronary artery abnormality. Two comparisons, CRP (p = 0.147, I2 = 52.3%) and ferritin (p = 0.128, I2 = 56.9%) levels, showed a moderate degree of heterogeneity. Rest of the comparisons showed no significant differences in statistical heterogeneity.
KD during SARS-CoV-2 epidemic vs. before SARS-CoV-2 epidemic22,30−32
Patients with KD during SARS-CoV-2 epidemic had lower levels of ALC (× 109/L) [WMD (95%CI): -2.59 (-2.82, -2.37), p < 0.001] and higher levels of CRP (mg/dL) [WMD (95%CI): 4.59 (2.33, 6.86), p < 0.001], compared to patients with KD before SARS-CoV-2 epidemic. There were no significant differences in statistical heterogeneity.
MIS-C in different age groups [younger age (0–5 years) vs. medium age (6–12 years) vs. older age(13–20 years)]28,29
MIS-C patients in younger age group had lower levels of ferritin (ng/mL) [WMD (95%CI): -285.78 (-457.04, -114.53), p < 0.01] than those in medium age group.
MIS-C patients in younger age group had lower CRP(mg/L) [WMD(95%CI): -88.75 (-122.67, -54.84), p < 0.001] and mildly higher D-dimer (µg/mL)[WMD(95%CI): 1.49 (0.37, 2.61), p < 0.01] levels than those in older age, but had same levels of ESR (mm/hr) [WMD (95%CI): -7.79 (-16.38, 0.80), p = 0.076].
MIS-C patients in medium and older age groups had same levels of CRP (mg/L) [WMD (95%CI): -24.95 (-58.96, 9.06), p = 0.150] and ferritin (ng/mL) [WMD (95%CI): -91.69 (-413.85, 230.46), p = 0.577]. The comparison with ferritin levels (p = 0.148, I2 = 52.2%) showed a moderate degree of heterogeneity.
Sensitivity analysis and publication bias
The results from sensitivity analysis were fairly similar and verified the stability of our analytical models. In addition, results from both the models [random effects model (REM) and fixed effects model (FEM)] were consistent, which indicated reliability in interpreting the combined results. Since the number of included studies in each comparison group was less than 10, we did not assess for publication bias.